Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) #305

yantosca · 2024-02-12T21:31:53Z

Name and Institution (Required)

Name: Bob Yantosca
Institution: Harvard + GCST

Confirm you have reviewed the following documentation

Contributing guidelines

Describe the update

Due to legacy, there were several benchmark-related files in the gcpy/ folder. These have now been abstracted into the gcpy/benchmark/modules/ folder (along with several YAML configuration files). Imports and function calls in the various benchmark routines have been updated accordingly.

Expected changes

Prior to this update, the gcpy/gcpy/ folder contained these files:

aod_species.yml           emission_inventories.yml       __pycache__/
append_grid_corners.py    emission_species.yml           raveller_1D.py
benchmark/                examples/                      regrid.py
benchmark_categories.yml  file_regrid.py                 regrid_restart_file.py
benchmark_funcs.py        grid.py                        species_database.yml
bpch_to_nc_names.yml      grid_stretching_transforms.py  ste_flux.py
budget_ox.py              __init__.py                    units.py
budget_tt.py              lumped_species.yml             util.py
constants.py              mean_oh_from_logs.py           _version.py
cstools.py                oh_metrics.py
date_time.py              plot/

after this update, it now contains these files:

append_grid_corners.py  date_time.py                   plot/                   units.py
benchmark/              examples/                      plot.py                 util.py
bpch_to_nc_names.yml    file_regrid.py                 __pycache__/            _version.py
colormaps/              grid.py                        raveller_1D.py
constants.py            grid_stretching_transforms.py  regrid.py
cstools.py              __init__.py                    regrid_restart_file.py

and the gcpy/benchmark/modules folder now contains these files:

aod_species.yml                budget_ox.py               oh_metrics.py
benchmark_categories.yml       budget_tt.py               __pycache__/
benchmark_drydep.py            emission_inventories.yml   README.md
benchmark_funcs.py             emission_species.yml       run_1yr_fullchem_benchmark.py
benchmark_models_vs_obs.py     GC_72_vertical_levels.csv  run_1yr_tt_benchmark.py
benchmark_models_vs_sondes.py  __init__.py                species_database.yml
benchmark_utils.py             lumped_species.yml         ste_flux.py

Furthermore, several benchmark-related functions have been moved out of gcpy/util.py and into gcpy/benchmark/modules/benchmark_utils.py. This helps to keep the general GCPy routines separate from the benchmark-specific routines.

Also, the functionality to rename variables starting with SpeciesConc_ to SpeciesConcVV_ has been moved into a the rename_speciesconc_to_speciesconcvv function in gcpy/benchmark/modules/benchmark_utils.py.

Lastly, we now have fixed a few items from #288, namely:

The YAML configuration file is now copied to each of the benchmark results folders (i.e. BenchmarkResults/GCC_version_comparison, BenchmarkResults/GCHP_GCC_comparison, BenchmarkResults/GCHP_version_comparison), not just for the TransportTracers benchmark, but for all benchmark types.
In TransportTracers benchmarks:
- Mass conservation tables are created for Ref and Dev
- Mass tables are created for Ref and Dev
- Radionuclude budget tables are created for Ref and Dev
- STE flux tables are created for Ref and Dev (GCC vs GCC only)
A bug was fixed where the directory was used in several tables instead of the version string. This is now fixed.

Tagging @lizziel.

Related Github Issue(s)

gcpy/benchmark/benchmark_slurm.sh - "${config/.yml/.log}: -> "${config/.yml/.log}" Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/aod_species.yml gcpy/benchmark_categories.yml gcpy/benchmark_funcs.py gcpy/budget_ox.py gcpy/budget_tt.py gcpy/emission_inventories.yml gcpy/emission_species.yml gcpy/lumped_species.yml gcpy/oh_metrics.[py gcpy/species_database.yml gcpy/ste_flux.py - Moved from gcpy -> gcpy/benchmark/modules gcpy/benchmark/run_benchmark.py gcpy/benchmark/modules/run_1yr_fullchem_benchmark.py gcpy/benchmark/modules/run_1yr_tt_benchmark.py - Updated import statements accordingly - Import benchmark routines by name gcpy/mean_oh_from_logs.py - Removed, this is obsolete gcpy/__init__.py - Removed references to benchmark_funcs, budget_ox, budget_tt, mean_oh_from_logs, oh_metrics, and ste_flux CHANGELOG.md - Updated accordingly

The following benchmark-related functions were moved from gcpy/util.py to gcpy/benchmark/modules/benchmark_utils.py: - get_species_categories - archive_species_categories - get_lumped_species_definitions - archive_lumped_species_definitions - add_lumped_species_to_dataset Also: gcpy/benchmark/modules/benchmark_util.py - Now import xarray - Add constants for YAML files that are now contained in the gcpy/benchmark/modules folder. These should not change very much, except when benchmark categories or lumped species change. - Updated Pydoc headers gcpy/benchmark/modules/benchmark_funcs.py - Now import several constants and functions from benchmark_utils.py - Removed local constants for YAML files - Updated function calls accordingly to be consistent with the new functions in benchmark_utils.py CHANGELOG.md - Updated accordingly Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/benchmark/modules/benchmark_utils.py - Now use "ofile" instead of "lumped_spc" in print statements in these functions: - archive_lumped_species_definitions - archive_species_categoriews Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/util.py - Removed "import shutil" statement gcpy/benchmark/modules/benchmark_utils.py - Added "import shutil" statement. This is necessary to use the shutil.copyfile command Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

…file docs/requirements.txt - Symbolic link to docs/environment_files/read_the_docs_requirements.txt. The online RTD builds looks for a requirements.txt file. CHANGELOG.md - Updated accordingly Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/benchmark/modules/benchmark_utils.py - Added function "rename_speciesconc_to_speciesconcvv", to centralize the renaming of variables starting with "SpeciesConc_" to "SpeciesConcVV_". This is necessary for backwards compatibility with GEOS-Chem versions prior to 14.1.0. gcpy/benchmark/modules/benchmark_funcs.py gcpy/benchmark/modules/benchmark_models_vs_obs.py gcpy/benchmark/modules/benchmark_models_vs_sondes.py gcpy/benchmark/modules/budget_tt.py - Now import and call "rename_speciesconc_to_speciesconcvv" from benchmark_utils.py CHANGELOG.md - Updated accordingly Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/benchmark/modules/budget_tt.py - Add quiet=True to read_config_file to suppress printout of the species_database.yml path - Add the version number (globvars.devstr) to the file path, so we can save Ref & Dev radionuclide budget tables gcpy/benchmark/modules/ste_flux.py - Add the version number (globvars.devstr) to the file path, so we can save Ref & Dev STE flux tables gcpy/benchmark/modules/benchmark_funcs.py - Updated Pydoc headers - Add the version number to the output file name in routine make_benchmark_mass_conservation_table gcpy/benchmark/modules/run_1yr_tt_benchmark.py - Updated function calls to transport_tracers_budgets, make_benchmark_ste_tables, and make_benchmark_massconv_tables so that we can now print Ref & Dev tables for GCC vs GCC, GCHP vs GCC, and GCHP vs GCHP comparisons. CHANGELOG.md - Updated accordingly Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/util.py - Add new function, copy_file_to_dir, which is a wrapper for the shutil.copyfile function. This abstracts the use of shutil.copyfile to a single location. gcpy/benchmark/modules/run_1yr_fullchem_benchmark.py gcpy/benchmark/modules/run_1yr_tt_benchmark.py gcpy/benchmark/run_benchmark.py - Updated Pydoc header documentation w/ more up-to-date usage information - Import copy_file_to_dir from gcpy/util.py - No longer import shutil.copyfile - Call copy_file_to_dir to copy __file__ (i.e. the name of the scripb being run) to the benchmark results directories - Call copy_file_to_dir to copy the YAML configuration file to the benchmark results directories Also in gcpy/benchmark/run_benchmark.py: - In main: Add the configuration file name (the 1st argument) to the config dict (as config["configuration_file_name"]) so that we can use it to copy the config file to benchmark results directories. to the benchmark run folders CHANGELOG.md - Updated accordingly Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/benchmark/modules/benchmark_utils.py - Now call util.copy_file_to_dir instead of shutil.copyfile in the following functions: - archive_lumped_species_definitions - archive_species_categories Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

gcpy/benchmark/modules/run_1yr_fullchem_benchmark.py

gcpy/benchmark/modules/run_1yr_tt_benchmark.py

msulprizio · 2024-03-06T18:26:35Z

gcpy/benchmark/modules/run_1yr_tt_benchmark.py

+    #bmk_mons_gchp_ref = all_months_gchp_ref[bmk_mon_inds]
+    #bmk_sec_per_month_ref = sec_per_month_ref[bmk_mon_inds]


Can these lines be removed?

@msulprizio: Now fixed in commit 3b380be.

gcpy/benchmark/modules/run_1yr_tt_benchmark.py

gcpy/benchmark/modules/run_1yr_fullchem_benchmark.py - Fixed incorrect comment in Pydoc header gcpy/benchmark/modules/run_1yr_tt_benchmark.py - Bug fix: gchp_vs_gchp_refrstdir should use ["data"]["ref"] instead of ["data"]["dev"] search keys - Removed commented-out code that defines bmk_mons_gchp_{ref,dev} and bmk_sec_per_month_{ref,dev} variables, these aren't used.

lizziel

Approved!

yantosca added 3 commits February 12, 2024 14:28

Fix typo in gcpy/benchmark/benchmark_slurm.sh script

90dd5e5

gcpy/benchmark/benchmark_slurm.sh - "${config/.yml/.log}: -> "${config/.yml/.log}" Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

yantosca added category: Feature Request New feature or request topic: Benchmark Plots and Tables Issues pertaining to generating plots/tables from benchmark output topic: Structural Modifications Related to GCPy structural modifications (as opposed to scientific updates) labels Feb 12, 2024

yantosca self-assigned this Feb 12, 2024

yantosca added 6 commits February 12, 2024 16:42

yantosca linked an issue Feb 21, 2024 that may be closed by this pull request

[FEATURE REQUEST] Transport tracer benchmark improvements #288

Open

yantosca requested a review from lizziel February 21, 2024 21:45

yantosca marked this pull request as ready for review February 21, 2024 21:45

yantosca changed the title ~~[WIP] Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder~~ Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) Feb 21, 2024

yantosca mentioned this pull request Feb 21, 2024

[FEATURE REQUEST] Transport tracer benchmark improvements #288

Open

yantosca requested a review from msulprizio March 6, 2024 15:21

msulprizio reviewed Mar 6, 2024

View reviewed changes

yantosca mentioned this pull request Mar 6, 2024

Combine Ref and Dev in the same file for TransportTracers mass conservation tables #308

Closed

1 task

yantosca requested a review from msulprizio March 6, 2024 19:01

lizziel approved these changes Mar 7, 2024

View reviewed changes

msulprizio approved these changes Mar 8, 2024

View reviewed changes

yantosca merged commit f7d0dc5 into dev Mar 8, 2024

yantosca deleted the feature/move-bmk-code-to-gcpy-benchmark-folder branch March 8, 2024 18:19

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) #305

Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) #305

Uh oh!

yantosca commented Feb 12, 2024 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

msulprizio Mar 6, 2024

Uh oh!

yantosca Mar 6, 2024

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lizziel left a comment

Uh oh!

Uh oh!

		#bmk_mons_gchp_ref = all_months_gchp_ref[bmk_mon_inds]
		#bmk_sec_per_month_ref = sec_per_month_ref[bmk_mon_inds]

Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) #305

Move remaining benchmark-related code out of gcpy and into gcpy/benchmark/modules folder (plus other improvements) #305

Uh oh!

Conversation

yantosca commented Feb 12, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Name and Institution (Required)

Confirm you have reviewed the following documentation

Describe the update

Expected changes

Related Github Issue(s)

Uh oh!

Uh oh!

Uh oh!

msulprizio Mar 6, 2024

Choose a reason for hiding this comment

Uh oh!

yantosca Mar 6, 2024

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lizziel left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

yantosca commented Feb 12, 2024 •

edited

Loading